Aggression, social cognitions, anger

and sadness in bullies and victims

Marina Camodeca
1,2
and Frits A. Goossens
1
1
Free University, Amsterdam, The Netherlands;
2
Pedological Institute, Amsterdam, The Netherlands
Background: The present study aimed to investigate childrens social information processing (SIP) and
emotions in the bullying situation, taking into account reactive and proactive aggression. More spe-
cically, we investigated the way in which children interpret social information, which goals they select,
how they evaluate their responses and which emotions they express in hypothetical situa-
tions. Method: The participants comprised 242 Dutch children (120 girls and 122 boys; mean age:
117.2 months), who were assigned by means of peer nominations (Salmivalli, Lagerspetz, et al., 1996) to
one of the following roles: bully (n 21), follower of the bully (n 38), victim (n 35), defender of the
victim (n 48), outsider (n 52) and not involved (n 32). Sixteen children (including 3 bully/victims)
were not given any role. The reactive and proactive aggression scale (Dodge, & Coie, 1987) was lled out
by teachers in order to test the association between these types of aggression and involvement in
bullying. Children were presented with ambiguous scenarios and responded to questions about attri-
bution of intent, goal selection and emotions (anger and sadness). In addition, two questionnaires were
administered to children: one assessed perceived self-efcacy in performing aggression, inhibiting
aggression and using verbal persuasion skills, and the other assessed expected outcomes from beha-
ving aggressively or prosocially. Results: Results showed that while reactive aggression was common
in bullies and victims, proactive aggression was only characteristic of bullies. Both bullies and victims,
compared to the other children, scored higher on hostile interpretation, anger, retaliation and ease of
aggression. Bullies and followers claimed that it was easy for them to use verbal persuasion, while
victims turned out to be the saddest group. All children, irrespective of their role in the peer group,
thought that aggressive as well as prosocial behavior was more likely to produce desired results from a
friendly peer than from an aggressive one. Conclusions: Bullies and victims seem to be similar in
reactive aggression, SIP, and in the expression of anger, but the motivations which lead to their behavior
may be different, as well as the nal outcomes of their acts. Keywords: Bullying, victimization,
reactive aggression, social information processing, emotions. Abbreviations: PRS: Participant Role
Scale; RePro: Reactive and Proactive Aggression Questionnaire; SIP: social information processing.
Bullying is a phenomenon characterized by negative
actions towards a peer, with the intention to hurt
(Olweus, 1991, 1993). The actions of the bully are
repeated over time and may include physical or
verbal aggression (Olweus, 1993; Boulton, & Un-
derwood, 1992), and relational harassment (Bjorkq-
vist, Lagerspetz, & Kaukiainen, 1992; Crick, &
Grotpeter, 1995; Wolke, Woods, Bloomeld, & Kar-
stadt, 2000), which harms others by means of social
manipulation, social exclusion, and malicious ru-
mors. There is usually an imbalance of power be-
tween the bullies and their victims. Bullying takes
place within relatively small and stable settings (like
classes), which are characterized by the presence of
the same people (e.g., children). Generally, children
other than the bullies and their victims are also in-
volved in the bullying process and may actually
maintain the bullying by supporting the bully or
failing to defend the victim. Salmivalli, Lagerspetz,
Bjorkqvist, O

sterman, and Kaukiainen (1996) sug-

gested that all the children in a particular class play
a role in bullying and that only few of them may be
considered to be uninvolved.
Many studies have investigated the character-
istics of children involved in bullying (Boulton, &
Smith, 1994; Hawker, & Boulton, 2000; Pellegrini,
Bartini, & Brooks, 1999; Smith, & Brain, 2000;
Smith et al., 1999), but not much is known about
the social cognitions and emotions of bullies and
victims. How do bullies and victims encode and
interpret social cues? How do they respond to
them? Are they really all that different? The general
goal of this study was to apply the social informa-
tion processing (SIP) approach (Crick, & Dodge,
1994; Dodge, 1986) to the bullying phenomenon in
order to investigate how bullies and victims read
social information and how they react to it. The role
of emotion was also investigated. Furthermore, we
made a distinction between reactive and proactive
aggression to explain the relationship between
bullying and SIP, using as a link those studies
which investigated the SIP of reactively and proac-
tively aggressive children (Crick, & Dodge, 1996;
Dodge, & Coie, 1987; Dodge, Lochman, Harnish,
Bates, & Pettit, 1997).
Social Information Processing (SIP)
The SIP theory was reformulated by Crick and
Dodge (1994), after being initially presented by
Journal of Child Psychology and Psychiatry 46:2 (2005), pp 186197 doi: 10.1111/j.1469-7610.2004.00347.x
Association for Child Psychology and Psychiatry, 2004.
Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA
Dodge (1986). It postulates that social information
processing takes place in ve mental steps, in a
circular formula, leading to a nal behavioral en-
actment (step 6). In step 1, children (or people in
general) code social cues from the environment and
focus on those that are more important through
selective attention. These cues are then given
meaning (step 2) through interpretation of others
intentions and causal attributions. In step 3, chil-
dren clarify their goals, i.e., what they want to
achieve in order to produce particular outcomes. In
step 4 children search for possible responses from
long-term memory and are inuenced by the attri-
butions they have made and by the goals they want
to achieve. In step 5 children choose one of these
responses by considering the content of the re-
sponse itself, the outcome they expect from it, and
their self-efcacy in performing it. Finally, in step 6,
they enact the behavior chosen. This, in its turn,
requires monitoring, during which attention to new
cues to be encoded is employed. After step 6 the
cycle starts again (Dodge, & Crick, 1990), because
although individuals are engaged in parallel pro-
cesses at the same time, the single stimulus follows
a linear sequence in which feedback loops are
possible across the steps (Crick, & Dodge, 1994;
Pakaslahti, 2000).
Processing the whole cycle in a skillful way
leads to social competence, while biased processing
may lead to aggression and social deviance (Crick,
& Dodge, 1996; Dodge et al., 1997; Pettit, Polaha,
& Mize, 2001; Zelli, Dodge, Lochman, Laird, &
Conduct Problems Prevention Research Group,
1999). Aggressive children encode fewer and less
benign social cues, because of memory decits or
selective attention (step 1), attribute more hostile
intentions (step 2), select goals which damage the
relationship (step 3), generate fewer prosocial re-
sponses (step 4), evaluate aggressive responses
more favorably, expect positive outcomes from ag-
gressive behavior, and feel more self-condent in
performing it (step 5). Finally, this process leads to
the enacting of aggressive behavior (step 6) (Dodge,
& Crick, 1990; Quiggle, Garber, Panak, & Dodge,
1992).
We used the SIP approach for several
reasons. First of all, it has been applied very suc-
cessfully to explain the thinking of aggressive
children (Crick, & Dodge, 1999; Pakaslahti, 2000;
Pettit et al., 2001). Second, it makes it possible to
study different aspects (steps) of social process-
ing as well as the links between early decits and
later performance. Third, on the basis of SIP re-
search, a set of tested materials have become
available for measuring the various steps (Camo-
deca, Goossens, Schuengel, & Meerum Terwogt,
2003; Crick, & Dodge, 1994; Orobio de Castro,
2000; Quiggle et al., 1992; Perry, Perry, & Ras-
mussen, 1986) and these can be applied in
research on bullying.
Emotion
Crick and Dodge (1994, 1999) have indicated that
the SIP framework would be enhanced by consider-
ing the role of emotion as well. Lemerise and Arsenio
(2000) integrated emotions and cognitions in the
model and claimed that all the steps in the process
are affected by emotion. Encoding and interpretation
of social cues (steps 1 and 2 of SIP) can be inuenced
by anger, mood or type of relationship with the per-
petrator; selection of goals (step 3) by anger or em-
pathy with the victim; response generation and
decision (steps 4 and 5) by pre-existing emotions,
representation of past experiences, or capacity to
regulate emotions; nal enactment (step 6) by emo-
tion control or capacity to read and convey emotions.
Behavior is inuenced by emotion, which may
arise from thoughts (Graham, & Hoehn, 1995; We-
iner, 1995): cognitive and emotional factors are ()
interrelated dimensions of the same reaction or
process (Loeber, &Coie, 2001, p. 395). Children who
are extremely intense in their experience and ex-
pression of emotion (emotionality) and who present
poor emotion regulation skills have been found to be
at risk of problem behaviors and social maladjust-
ment (Eisenberg, & Fabes, 1992; Loeber, & Coie,
2001; Murphy, & Eisenberg, 1997; Pakaslahti,
2000; Pettit et al., 2001). Negative emotions, which
were found in bullies and victims (Karatzias, Power,
& Swanson, 2002), decrease the likelihood of ade-
quate behavioral responses and restrict cognitive
capacity to solve problems (Pakaslahti, 2000). Anger
is also a characteristic of reactively aggressive chil-
dren (Dodge et al., 1997; Dodge, & Coie, 1987). Or-
obio de Castro (2000) found that antisocial boys said
they became angrier and less sad when provoked,
and mentioned fewer adaptive emotion regulation
strategies than the control group. In view of these
ndings we included emotions in our study.
Reactive and proactive aggression
One of the most accepted distinctions between dif-
ferent types of aggression is the one between reactive
and proactive aggression (Dodge, 1991; Dodge, &
Coie, 1987; Pellegrini et al., 1999; Price, & Dodge,
1989; Pulkkinen, 1996; Salmivalli, & Nieminen,
2002). Reactive aggression is a defensive response to
provocation or trouble, a way to defend oneself and
to retaliate against abuse, and is accompanied by
anger. Proactive aggression, on the other hand, is a
goal-directed, deliberate and cold-blooded action,
useful to achieve goals, offensive and provocative,
requires no stimulus and may be characterized by
pleasure or satisfaction (Boulton, & Smith, 1994;
Crick, & Dodge, 1996; Roland, & Idse, 2001).
Are bullies and victims reactively and/or proac-
tively aggressive? Crick and Dodge (1999) and Price
and Dodge (1989) advanced the hypothesis that bul-
lies are proactively aggressive, while Kochenderfer
Bullying, emotion and social cognition 187
and Ladd (1997) reported that victims display re-
active aggression. However, recent studies have
found that bullies show both reactive and proactive
aggressive behavior, while victims only show reactive
aggression (Camodeca, Goossens, Meerum Terwogt,
& Schuengel, 2002; Pellegrini et al., 1999; Roland, &
Idse, 2001; Salmivalli, & Nieminen, 2002).
Crick and Dodge (1996) investigated differences in
social information processing and reported that
reactively aggressive children attribute hostile intent
to their peers and respond in an aggressive way.
Thus, they present decits in interpreting social
cues. Proactively aggressive children evaluate
aggression in a more positive way and as a valid
means to reach goals. They differ from the others in
goal clarication. Similar outcomes have been
reported by Dodge and Coie (1987) and Dodge
(1991).
Research questions and hypotheses
The present work was aimed at investigating differ-
ences and similarities among bullies, victims and the
other children in respect to types of aggression, SIP
and regulation of emotions. Recently, Karatzias et al.
(2002) advanced the hypothesis that when treated as
a single group, bullies and victims differ from those
not involved in respect to well-being, quality of
school life and certain personality factors, and were
thus more similar than is usually thought. Some
authors (Camodeca et al., 2002; Pellegrini et al.,
1999; Roland, & Idse, 2001; Salmivalli, & Niemi-
nen, 2002) also found bullies and victims to be
similar in respect to aggression. More specically,
both bullies and victims were reactively aggressive,
i.e., they responded with anger to provocation and
used aggression to defend themselves, while only
bullies were proactively aggressive, i.e., they used
aggression to harass and provoke. We expected to
nd the same results in this study.
We surmise that if a child is reactively aggressive
and shows decits at the beginning of SIP (as Crick,
& Dodge, 1996, claimed), these decits continue
throughout the whole process and inuence every
step. If this is the case and if bullies and victims are
both reactively aggressive, we may suppose that
bullies and victims encode, interpret and react to
social cues in a similar way and that they express the
same emotions. But if bullies are also proactively
aggressive, they may also present biases in the nal
steps of SIP (Crick, & Dodge, 1996) and their moti-
vation for behaving aggressively may be different
from that of the victims. For this reason differences
between bullies and victims can still be found.
More specically, in this study we expected both
bullies and victims to display reactive aggression, to
encode cues (step 1) in a biased way, to misinterpret
ambiguous situations (step 2) and to interpret intent
as hostile (Menesini, 1999; Quiggle et al., 1992). As a
consequence of this, they may select antisocial goals
(step 3) and respond with counter-aggression. This
expectation is based on the nding that it is the
perception of the intention which determines the
behavioral response rather than the intention itself
(Dodge, Murphy, & Buchsbaum, 1984). For this
reason, we expected both bullies and victims to claim
it was easy for them to react aggressively (step 5).
Conversely, we expected bullies and victims to en-
counter greater difculty in inhibiting their own ag-
gression.
Bullies were expected to think that aggression
would produce tangible rewards and reduce aver-
sive treatment (step 5) more often than all the
other children (Crick, & Dodge, 1994; Dodge et al.,
1997; Perry et al., 1986). The peer who was the
target of the behavior was also taken into consid-
eration, as it was found that less favorable out-
comes were expected when interacting with an
aggressive peer than with a nonaggressive one
(Perry et al., 1986). We did not investigate step 4
for practical reasons, as we considered it too time-
consuming to interview these relatively young
children about how they thought they would re-
spond, as they already had to answer so many
questions. Furthermore, we had investigated re-
sponse construction in a previous paper (Camo-
deca et al., 2003). We did not investigate step 6
either, since we were interested in the mental
process and not in the ultimate behavior.
As for emotions, we expected both bullies and
victims to respond with anger in the face of adversity,
as reactively aggressive children do (Arsenio, &
Lemerise, 2001; Dodge, & Coie, 1987; Dodge et al.,
1997). Moreover, we expected victims to show more
sadness than their classmates when confronted with
an unpleasant situation (Boulton, & Underwood,
1992; Quiggle et al. 1992).
To sum up, the purpose of the present study was
to investigate similarities and differences between
bullies and victims in displaying aggression, in SIP
and in the emotions of sadness and anger, taking
into account multi-informant measures and different
roles in bullying.
Method
Sample
The subjects were 242 Dutch children (120 girls and
122 boys) with a mean age of 117.2 months (SD 8.4),
from the fth (49.6%) and sixth grade (50.4%) of four
elementary schools in the Netherlands. The pupils came
from various socioeconomic backgrounds, but the
majority were from middle class families. Fewer than
5% were of non-Dutch origin. Parents were asked by
letter to consent to their childrens participation in the
study. The response rate was high (over 90%). The pu-
pils involved in this particular study are part of a lon-
gitudinal study and were also tested using other
measures at different points in time (Camodeca et al.,
2002, 2003).
188 Marina Camodeca and Frits A. Goossens
Procedure
Bullying and victimization were operationalized
through a peer report measure, the Participant Role
Scales (Salmivalli, Lagerspetz, et al., 1996), as children
are the best informants about bullying. Another reason
for choosing this instrument was that it has been
developed to assess the different roles children may
play in the bullying process, and not only those of
bully and victim. Pupils were tested individually by
trained students in a separate room. They were told
that all information would be treated as condential
and that it was better not to discuss what they had
said with their peers.
A questionnaire to test reactive and proactive
aggression was administered to teachers, who had to
rate each child on different items.
Four scenarios were administered individually to
children to test the processing of social information
from the rst to the third step of the model, and to test
the emotions of anger and sadness. One questionnaire
was used to assess self-efcacy and another one to
assess expected outcomes; they tapped the SIP step of
response evaluation. The scenarios and the two ques-
tionnaires were administered to the group in the
classroom.
In this way we could take advantage of a peer report
measure, a teacher report measure, and three self-
report measures.
Measures
The Participant Role Scales (PRS). The PRS was
designed by Salmivalli, Lagerspetz, et al. (1996) to
assess roles in the bullying situation. The original
questionnaire consisted of 50 descriptors of behavior
according to which the children were asked to rate each
of their peers and themselves on a 3-point scale. Five
scales were constructed: bully, reinforcer of the bully,
assistant of the bully, defender of the victim, and out-
sider. In order to identify the victims the authors used
nominations.
Oude Nijhuis (2001) validated the PRS in the Neth-
erlands. Instead of the original rating procedure used
by Salmivalli, Lagerspetz, et al. (1996), she employed a
nomination procedure in which children nominated one
or more classmates for each descriptor. This was found
to enhance data collection (cf. Sutton, & Smith, 1999;
Sutton, Smith, & Swettenham, 1999b). Further, Oude
Nijhuis (2001) added seven victimization items taken
from the Aggression and Victimization Scale (Perry,
Kusel, & Perry, 1988). In this way a new scale for
victims was created and used instead of the original
nomination procedure. Using this format, Oude Nijhuis
(2001) found that defenders were the most popular,
bullies turned out to be rejected or controversial,
while victims were mostly neglected. Outsiders
predominantly had average status.
We kept the changes made to the PRS by Oude
Nijhuis (nomination procedure and victimization items)
and in addition we deleted items with the lowest load-
ings on the factors. The new PRS now consisted of 32
items, which were subjected to a Principal Component
Analysis (PCA) with varimax rotation, yielding four
factors. The total variance explained was 74.6%. Four
items were deleted because of low loadings (< .50) or
cross-loadings; the remaining items loaded .50 or
higher on one dimension. All the bullying items (bully,
assistant and reinforcer) loaded on the rst factor,
which was also found by Salmivalli, Lappalainen, and
Lagerspetz (1998) and by Sutton and Smith (1999). But,
despite the high correlation among the items for bully,
assistant and reinforcer, these authors kept these roles
separate on the basis of their content. In the present
study, we created a scale for bully and a scale for fol-
lower (merging the items for assistant and reinforcer, cf.
Sutton et al., 1999b).
The items for defender, outsider and victim loaded on
three separate factors, which were kept on separate
scales. The number of items and the reliability coef-
cients for the ve scales were as follows: bully (6 items,
a .97), follower of the bully (reinforcers plus assist-
ants; 8 items, a .93), outsider (6 items, a .91),
defender (4 items, a .85), victim (4 items, a .91).
Salmivalli, Lagerspetz, et al. (1996) reported reliabilities
ranging from .81 to .93, and Sutton et al. (1999b) also
reported reasonably high reliabilities (ranging from .55
to .88). Next, the scores were standardized by class
using z-scores. We used the procedure employed by
Salmivalli, Lagerspetz, et al. (1996) to assign each child
a role. A child was assigned a role if the score on the
scale designed to assess that role was above the mean
and if the difference between this scale score and the
next highest scale score was at least .1. Pupils who re-
ceived almost equal scores on two or more scales (i.e.,
whose difference was lower than .1, n 13; 5.4%) were
considered as not having a clear role and were not
included in the analyses. We made an exception only for
the scales of bully and follower, in which case we
assigned the role in which the score was highest, even
when the difference between the two scores was smaller
than .1. We found also 3 bully/victims (1.2%), being
those children whose scores on the bully or the follower
scale and on the victimscale were higher than the mean,
while the difference between the pro-bully and the victim
scores was smaller than .1; scores on the defender and
outsider scales were lower than the mean. However, we
decided to remove these 3 children from the sample as
too few to be included as a separate group.
Ultimately, the sample employed in the analyses
consisted of 226 children (108 girls and 118 boys). Of
these, 21 (9.3%) pupils were assigned the role of bully,
38 (16.8%) the role of follower, 52 (23%) the role of out-
sider, 48 (21.2%) the role of defender and 35 (15.5%) the
role of victim. Contrary to Salmivalli, Lagerspetz, et al.
(1996), who also labeled those who scored below the
mean on all scales as not having a clear role, we con-
sidered these children as not involved (n 32; 14.2%) in
the bullying situation. The incidence of children in each
role and the gender distribution are comparable to other
studies which used the same instrument (Menesini, &
Gini, 2000; Salmivalli et al., 1998; Salmivalli, Huttunen,
&Lagerspetz, 1997; Salmivalli, Lagerspetz, et al., 1996).
In addition, in our study bullies and followers were more
often boys, while defenders and outsiders were more
often girls (v
2
42.13 (5); p < .001).
The Reactive and Proactive Aggression Question-
naire (RePro). The RePro was developed by Dodge
and Coie (1987). The original questionnaire consisted
Bullying, emotion and social cognition 189
of 12 aggressive behavior items for reactive (e.g., When
teased, strikes back) and proactive aggression (e.g.,
Uses physical violence to dominate), plus 12 ller
items. Later the authors decided to limit the two scales
to three items each, choosing only the items which
loaded highest on the factor analysis. We used 9 of the
original 12 aggression items (four for reactive aggres-
sion and ve for proactive aggression), plus 4 ller
items (which were excluded from the analysis). The
answer modality was a 7-point scale, instead of a 5-
point scale, as in the original version. Cronbachs al-
phas were .90 for the reactive aggression scale and .93
for the proactive aggression scale. We made use of
standardized scores within each class, in order to
minimize the effect of the raters.
Scenarios assessing steps 1, 2 and 3 of SIP and
emotions. Four stories (Orobio de Castro, 2000) were
presented to the children, in gender-appropriate ver-
sions, to investigate the rst three steps of SIP (encod-
ing of cues, interpretation of cues and clarication of
goals), and the emotions of anger and sadness. In these
stories a child is interacting with a peer when an
unpleasant incident is caused by the peer. Whether this
is intentional or not remains ambiguous. An example is:
Imagine you are taking turns on a computer game with
a classmate. When one is nished, its the others turn.
Now its your turn and you are doing well. You have
already reached the highest level, but you only have one
life left. You have never gotten as far as this, so you are
really doing your best. The other boy/girl is looking over
your shoulder. He/she sees how far you have got. Then
he/she says: Watch out! You have to be quick! and
pushes a button. But it was the wrong one, and now
youre dead.
In order to assess encoding of cues (step 1), children
were requested to tell the story again. The number of
essential elements (four per scenario: an outline of the
story, a report on the importance of the participants
goals, a description of the provocateurs behavior and a
description of its outcome) was counted to build the
variable encoding. Intercoder agreement among four
different raters (trained students) ranged from .66 to
.94. Differences in coding were solved by discussion.
To assess attribution of intent (step 2), children had to
answer the following questions: Do you think the other
child is mean? Do you think that he/she did it on pur-
pose? Do you think he/she is happy with what he/she
did? (Answers on a 3-point scale: No, I dont know and
Yes); How guilty do you think he/she is? (Answers on a
5-point scale from Not at all to Very much). These
questions, totaled across four scenarios, were subjected
to a PCA, and one factor (hostility) was extracted (16
items; 31.2% of variance explained; loadings higher
than .44; a .85).
At this point, interrupting the questions on the SIP
steps, we asked children three questions about emo-
tions: one for sadness (How sad would you feel?) and
two for anger (How angry would you feel if this hap-
pened to you? and How angry are you with him/her?).
All answers were on a 5-point scale (from Not at all to
Very much). We ran a separate PCA on sadness and
anger questions per four scenarios (57.0%and 49.2%of
variance explained for sadness and anger, loadings
higher than .70 and .60, respectively). Cronbachs
alphas were .75 for sadness (4 items) and .85 for anger
(8 items).
In order to assess childrens goals (step 3), we asked
pupils: How important is it for you1) to forget as soon
as possible? 2) to feel less angry? 3) to retaliate for what
he/she did? 4) to have a nice time together? 5) that the
other child does not feel guilty about what he/she did?.
Answers were on a 5-point scale, with higher scores
indicating more importance given. Factor analysis (PCA
with varimax rotation, 42.75% of variance explained)
revealed two factors: retaliation (item 3 per four stor-
ies 4 items; loadings higher than .78; a .82) and
prosocial goals (items 1, 2, 4 and 5 per four stories 16
items; loadings higher than .45; a .89).
Self-efcacy questionnaire assessing step 5 of
SIP. This questionnaire was used to assess response
decision by means of perceived ability to use aggression,
inhibition of aggression, and assertive behavior. It is a
shorter version (24 four-point items describing social
situations) of the questionnaire developed by Perry et al.
(1986). Children had to rate how easy it would be to
perform the behavior described in the item. High scores
indicate greater ease in performing the behavior. Two
versions, one for boys and one for girls, were employed.
Examples of items covering the different components
are as follows: Aggression: In the playground another
child bumps into you. Calling him/her bad names is ___
for you. Inhibition of aggression: One of your classmates
invites everyone to his/her party. You are not invited.
You would like to say something mean to him/her, but
decide not to. Not saying mean things to that child is ___
for you. Verbal persuasion: Some children want to play
a game you do not like. Proposing another game that you
like more is ___ for you. In the blanks children had to
indicate on a 4-point scale how easy it was for them to
perform the specied behavior in that situation by
circling one of the following possible answers:
DIFFICULT difcult easy EASY
The itemscores were subjected to a PCA with varimax
rotation (50.4% of variance explained). Four items were
deleted because they showed cross-loadings. All the
others loaded above .50 on one of the three expected
dimensions. Eight items loaded on aggression (a .90),
ve on inhibition of aggression (a .79) and seven on
verbal persuasion (a .80).
Expected outcomes questionnaire assessing step 5
of SIP. The expected outcome questionnaire was also
developed by Perry et al. (1986) and measures the de-
gree to which children are condent a certain outcome
will occur if they behave in a particular way in a given
situation. It was used to investigate the way in which
children decide upon their responses. We presented 16
situations to the children, 12 of which anticipated the
consequences of aggressive behavior and 4 of prosocial
behavior. These situations involved two different types
of target children, who were of the same sex as the
subjects. In eight situations (6 for aggressive behavior
and 2 for prosocial behavior) the target child was de-
scribed in advance as aggressive and mean (bossy, al-
ways wants to have his/her way, gets angry very easily,
hits people and calls them names), while in the other
eight (again, 6 and 2 for the two behaviors) the target
190 Marina Camodeca and Frits A. Goossens
child was described as friendly and nice (always
friendly, helpful, always knows nice things to do,
everyone wants to play with him/her). In this way we
could investigate whether the targets characteristics
inuenced the childrens expected outcomes. The two
types of target children were presented in a different
order, to minimize the effect of priming.
Examples of the situations are the following:
Aggression: (Target child) bullies you at school and
calls you names. You call him/her names too, in the
hope he/she will stop. What will he/she do now? Pro-
social behavior: One day you come to school with a
packet of crisps. (Target child) sees this packet of crisps
and tries to grab it from you. You want to push him/her
away, but decide to share the crisps with him/her. Do
you think (target child) will still try to get the crisps from
you? Children had to indicate how sure they were that
their behavior would succeed in stopping the others
behavior. Ratings were on a 4-point scale, with 1
meaning that the child was very sure that the conse-
quence would occur and 4 meaning that the child was
very sure that the consequence would not occur. A
factor analysis (PCA, varimax rotation; 34.0% of vari-
ance explained) pointed to the existence of two factors
with item loadings higher than .40. Alpha coefcients
were computed for the expected outcomes scales for
aggression (a .79) and prosocial behavior (a .52).
Then each factor was split into two according to the
items which were coupled to the aggressive target or to
the friendly target. In this way, each child had two
scores per scale: one for the expected outcomes with an
aggressive target (a .72 for aggression and a .54 for
prosocial behavior) and another for the expected out-
comes with a friendly target (a .65 for aggression and
a .25 for prosocial behavior).
Results
Scenarios: steps 1, 2 and 3 of SIP and emotions
A 2 (gender) 2 (grade) 6 (role) analysis of variance
with the six scales of the scenarios as dependent
variables (e.g., encoding, hostility, anger, sadness,
retaliation and prosocial goals) showed a trend effect
for grade (Pillais Trace .06; F (6, 195) 1.94;
p < .10), a signicant effect for role (Pillais
Trace .23; F (30, 995) 1.59; p < .05), for the in-
teraction between role and grade (Pillais
Trace .24; F (30, 995) 1.66; p < .05) and a trend
for the interaction between role and gender (Pillais
Trace .20; F (30, 995) 1.37; p < .10). Univariate
tests, means and standard deviations for the effect of
role are shown in Table 1. As for grade, the univari-
ate test did not reach signicance. The interaction
between role and grade was signicant for anger (F
(5) 2.35; p < .05), which also yielded a trend in the
interaction between role and gender. Bullies, de-
fenders and children not involved said they would be
angrier at grade 6 than at grade 5, compared to
victims, outsiders and followers who said they would
be angrier at grade 5. Boys said they would be angr-
ier than girls when they were bullies or defenders.
However, in absence of specic hypotheses, we did
not explore these differences any further.
Roles did not differ fromeach other in encoding. As
for the second step of SIP, both bullies and victims
attributed more hostile intentions to the perpetrator
in comparison to the other children. They all said
they would be angrier than the other children, while
victims said they would feel the saddest of the whole
sample (a slight difference was also found between
outsiders and followers, with the rst saying they felt
sadder). Victims wished to retaliate more often than
their classmates did, while bullies chose retaliation
more often than defenders. No differences were
found for selecting prosocial goals.
Self-efcacy questionnaire: step 5 of SIP
Three effects attained signicance when the MANO-
VA with the self-efcacy questionnaire as dependent
variable was used. These were the main effects of
gender (Pillais Trace .12; F (3, 198) 9.17;
p < .001), grade (Pillais Trace .04; F (3,
198) 2.96; p < .01) and role (Pillais Trace .13; F
(15, 600) 1.75; p < .05). Boys thought more often
than girls that it was easy to behave both aggressively
Table 1 Means, standard deviations (between parentheses) and tests of group differences of the raw scores of ambiguous scenarios
as a function of role in the bullying situation
Bully
(n 21)
Follower
(n 38)
Victim
(n 35)
Defender
(n 48)
Outsider
(n 52)
Not involved
(n 32)
F
(df 5)
Step 1
Encoding 13.0 (2.0) 13.1 (1.9) 13.2 (1.4) 13.7 (1.4) 12.9 (1.8) 13.2 (2.0) ns
Step 2
Hostility 19.1 (9.5)
y
a
15.2 (8.6)
y
b
19.3 (7.5)
a
14.7 (8.0)
b
15.7 (7.1)
y
b
13.8 (7.1)
b
3.47**
Emotions
Anger 21.2 (8.1)
a
16.9 (6.4)
b
20.3 (6.2)
y
a
17.0 (6.3)
b
17.8 (6.8)
y
b
16.1 (6.0)
b
3.13**
Sadness 7.8 (4.3)
bc
6.9 (3.6)
y
b
9.8 (3.4)
a
7.9 (3.2)
bc
8.2 (3.9)
y
c
7.3 (3.5)
bc
2.11

Step 3
Retaliation 5.5 (3.5)
ac
4.4 (4.2)
bc
6.3 (4.9)
a
3.3 (3.4)
b
4.2 (4.3)
bc
3.9 (3.3)
bc
3.57**
Prosocial 42.1 (13.2) 38.0 (13.0) 40.7 (12.0) 41.4 (12.5) 41.8 (12.4) 38.9 (11.7) ns
Note: Means in the same row with different subscripts (ac) differ signicantly at p < .05, two-tailed (

p < .10) by the least signicant

difference test.
ns non-signicant.

p < .10. *p < .05. **p < .01.
Bullying, emotion and social cognition 191
(F (1) 27.01; p < .001) and assertively (F (1) 7.88;
p < .01). Older children (sixth grade) were more self-
condent regarding verbal persuasion than the
younger ones (F (1) 5.48; p < .01). Univariate tests,
means and standard deviations for the effect of role
are shown in Table 2. Bullies, victims and followers
found it easier to behave aggressively than the other
children. Bullies and followers also claimed they
found it easy to persuade others verbally in compar-
ison to defenders and outsiders (bullies also had
higher scores than victims and children not involved).
No role differences were found for ease in inhibiting
aggression.
Expected outcomes questionnaire: step 5 of SIP
A multivariate analysis of variance was performed on
the scale scores, with three between-subject factors
(role in bullying, gender and grade) and one within-
subject factor (target: aggressive or friendly peer).
Three effects were signicant. The rst was the main
effect of grade (Pillais Trace .08; F (2, 198) 8.51;
p < .001; univariate test: F (1) 16.03; p < .001 for
aggression). Children in fth grade were less sure
than those in sixth grade that aggressive behavior
would be successful in reducing attacks from the
perpetrator or in obtaining rewards. The second
signicant effect was the interaction between gender
and grade (Pillais Trace .05; F (2, 198) 4.87;
p < .01; univariate test: F (1) 8.65; p < .01 for ag-
gression). While fth grade boys were less sure than
girls that they would obtain rewards by using ag-
gression, one year later it was the other way around.
The third signicant effect was the within-subject
effect (Pillais Trace .69; F (2, 198) 223.95;
p < .001). The univariate test of signicance (Huynh-
Feldt), means and standard deviations for the effect
of target are shown in Table 3. Children thought that
their behavior (either aggressive or prosocial) would
be less successful if they were interacting with an
aggressive child than if they were interacting with a
nonaggressive child.
Association between role in bullying
and reactive and proactive aggression
We investigated the link between involvement in
bullying on the one hand and reactive and proactive
aggression on the other by means of regressions.
Since we did not nd any effect due to different
grade, we excluded this variable from the analyses.
Gender was entered in the rst step, the two scores
for reactive and proactive aggression in the second
step, and the interactions between these and gender
in the third (the latter two steps entered using the
stepwise method). As the scores of the dependent
variables revealed some extreme cases, we ran the
regression analyses on the scores after they had
been normalized with the SPSS ranking program,
which centers the means. In addition, the variable
gender was centered (it was recoded into the values
of )1 and +1 for boys and girls, respectively). The
interaction between gender and the types of aggres-
sion did not yield any signicant results and was
therefore removed from the analyses, which were
rerun without the interaction effect. Table 4 shows
the results of the regressions obtained with the
ranking procedure.
The results support our expectation that bullying
is associated with both proactive and reactive
aggression, while victimization is associated only
with reactive aggression. Being a follower showed the
same pattern as being a bully (it was related to both
types of aggression), while being a defender was
negatively related to reactive aggression and being
an outsider was negatively related to both types of
aggression. Bullies and followers were more often
boys, while defenders, outsiders and victims were
more often girls.
Discussion
Some authors (Rubin, Bream, & Rose-Krasnor,
1991, p. 222) dened social competence as the
ability to achieve personal goals in social interaction
Table 2 Means, standard deviations (between parentheses) and tests of group differences of the self-efcacy questionnaire raw
scores as a function of role in the bullying situation
Bully
(n 22)
Follower
(n 38)
Victim
(n 35)
Defender
(n 48)
Outsider
(n 52)
Not involved
(n 32)
F
(df 5)
Aggression 23.2 (6.5)
a
23.9 (6.5)
a
21.8 (6.9)
ac
18.0 (5.3)
b
18.7 (6.0)
b
19.6 (6.0)
bc
2.14

Verbal persuasion 25.1 (2.1)

a
23.4 (3.6)
ac
21.9 (4.2)
bc
21.6 (4.9)
b
21.0 (4.9)
b
22.6 (3.7)
bc
2.27*
Inhibition 11.3 (3.6) 13.1 (4.0) 12.3 (3.6) 11.5 (3.6) 12.7 (4.0) 11.4 (3.7) ns
Note: Means in the same row with different subscripts (ac) differ signicantly at p < .05, two-tailed (

p < .10) by the least signicant

difference test.
ns non-signicant.

p < .10. *p < .05.

Table 3 Means, standard deviations (between parentheses)
and tests of group differences on the expected outcomes
questionnaire raw scores as a function of the target type
Aggressive
target
Friendly
target F (df)
Aggressive behavior 19.7 (3.1) 14.5 (3.0) 445.23 (1)***
Prosocial behavior 4.7 (1.6) 4.0 (1.2) 21.09 (1)***
***p < .001.
192 Marina Camodeca and Frits A. Goossens
while simultaneously maintaining positive relation-
ships with signicant others (cf. also Arsenio and
Lemerise, 2001). According to this view the bullies
behavior, although effective for the bullies them-
selves, is not socially competent as it does not take
into account the shared values of the group, social
norms and peers well-being. Our data seem to
support the view of Crick and Dodge (1994) who
claimed that the behavior of bullies is a result of a
processing bias and of decits in some stages of the
SIP. However, it would appear that bullies are not the
only socially inadequate subjects in our sample.
The results of the present study show that bullies
and victims display more decits in processing social
information than other children in the class, and that
they respond more emotionally to adverse conditions
(Lemerise, & Arsenio, 2000; Pellegrini et al., 1999).
We surmise that this similarity is due to their com-
mon reactive aggression. Indeed bullying was posit-
ively linked with reactive and proactive aggression,
while victimization was only associated with reactive
aggression. Bullies and victims are similar in their
reactive aggression, i.e., they defend themselves with
counter-aggression because they tend to perceive
threats in certain situations and therefore to respond
with anger and retaliation (Crick, & Dodge, 1994;
Graham, & Juvonen, 1998). However, their reactive
aggression, and therefore their way of processing
social information, may have different reasons and
motivations and produce different nal behaviors.
Our results on the SIP framework only partially
conrmed the claims of those studies which
investigated the SIP of reactively and proactively
aggressive children (Crick, & Dodge, 1996; Dodge
et al., 1997; Dodge, & Coie, 1987), namely that
reactively aggressive children (i.e., our victims and
bullies) present decits at the beginning of the SIP
cycle, while proactively aggressive children (i.e., our
bullies) show a different cognitive pattern in the nal
steps of the SIP cycle. In our studies, in fact, both
bullies and victims were found to present decits in
almost every step of the SIP, supporting the circular
formula of the SIP framework, in which every step
inuences the following one, as Crick and Dodge
(1994) themselves suggested. We can therefore argue
that if a child presents a cognitive bias when attrib-
uting intent, this is carried on along the whole
process, through selection of antisocial goals,
expression of anger, creation of aggressive
responses, feeling of self-efcacy in performing
aggression, and, possibly, in nal enactment of the
aggressive behavior. Bullies and victims interpret
ambiguous situations as hostile, failing in the sec-
ond step of SIP. As a consequence of their continued
exposure to bullying, victims do not trust others
(Champion, 2001; Smith, 1991), while bullies are so
used to harassing others on purpose that they think
everyone who behaves aggressively does so deliber-
ately. They also were similar in anger, which is an
emotion typical of reactively aggressive children
Table 4 Regression of the RePro on the PRS
Variable Predictors b R R
2
R
2
change
F(df)
Bully I step .45 .20 58.06 (1)***
Gender ).45***
II step .71 .50 .30 75.42 (3)***
Gender ).28***
Proactive .37***
Reactive .22*
Follower I step .49 .24 70.89 (1)***
Gender ).49***
II step .72 .52 .28 80.21 (3)***
Gender ).32***
Proactive .34***
Reactive .24*
Victim I step .02 .00 .10 (1)
Gender .02
II step .41 .17 .17 22.66 (2)***
Gender .18**
Reactive .44***
Defender I step . .43 .18 49.67 (1)***
Gender .43***
II step .54 .29 .11 46.78 (2)***
Gender .29***
Reactive ).36***
Outsider I step .41 .16 44.49 (1)***
Gender .41***
II step .66 .43 .27 56.75 (3)***
Gender .24***
Proactive ).28**
Reactive ).28**
*p < .05. **p < .01. ***p < .001.
Bullying, emotion and social cognition 193
(Dodge, 1991; Dodge, & Coie, 1987; Loeber, & Coie,
2001). Their anger follows from holding others re-
sponsible for negative actions against them (Camo-
deca et al., 2003; Graham, Hudley, & Williams,
1992; Weiner, 1995) and could therefore be consid-
ered an experience factor. But it also seems possible
that bullies and victims express anger as a conse-
quence of their hot-headed temperament (in this
case it would be anger leading to blame). The data do
seem to support Arsenio and Lemerises (2001) the-
sis that reactively aggressive children are easily
aroused and more likely to behave aggressively as a
result of outbursts of anger. As a consequence,
hostile attributions and anger may lead to retaliation
through the choice of aggressive goals (Crick, &
Dodge, 1994; Graham, & Juvonen, 1998; Loeber, &
Coie, 2001). In fact, another point of similarity be-
tween bullies and victims turned out to be their
choice of retaliation. Victims may resolve to select
goals which destroy the relationship either as a re-
sult of frustration and exasperation, or because they
are not capable of behaving prosocially, or because
they think that this is indeed a successful way of
defending themselves from the bullies attacks.
However retaliation and reactive aggression are
usually ineffective for them in gaining desired out-
comes (Pellegrini et al., 1999; Salmivalli, Karhunen,
& Lagerspetz, 1996; Schwartz, McFayden-Ketchum,
Dodge, Pettit, & Bates, 1998) and could indeed even
have the effect of making the bully more ruthless. On
the other hand, bullies, being also proactively ag-
gressive, may use retaliation as a further means to
reach their own goals, such as for instance obtaining
an object or achieving higher status in the peer
group. In fact, they may bully because they nd it
easy and useful for their purpose (Sutton, 2001;
Sutton, Smith, & Swettenham, 1999a).
Another similarity between bullies and victims is
that both of them claimed to nd it easy to behave
aggressively. This may be inuenced by anger, which
could increase condence in their capacity to re-
taliate aggressively. It can also be a consequence of
the previous SIP step: if a person selects antisocial
goals, it is likely that he/she feels condent of
achieving them. Finding it easy to use aggression is
in line with the role of bullies (Perry et al., 1986),
because they need to feel capable of displaying ag-
gression in order to obtain their goals. On the other
hand, although victims may think it is easy to be-
have aggressively, they are unlikely to be able to
defend themselves from attack in an effective way
(Egan, & Perry, 1998; Salmivalli, Karhunen, et al.,
1996). According to Paladino and Kochenderfer-
Ladd (2001), victims need to feel active and able
enough to cope with difcult situations.
In respect of two variables, bullies and victims
differed from each other. The rst was ease in using
verbal persuasion, which was a characteristic only of
bullies and followers. It might be possible that these
children know that they have dominant status
among their classmates and can use it to persuade
in other ways than aggression. However, studies by
Salmivalli (2001) and Salmivalli, Kaukiainen, Kai-
staniemi, and Lagerspetz (1999) found that bullies
self-esteem is not genuine, as bullies have a narcis-
sistic view of themselves, characterized by self-
aggrandizing tendencies, arrogance and dominance.
Also, Schippell and Vasey (2001) reported that
inated ratings of self-competence were related to
proactive aggression, which is typical only of bullies,
and not to reactive aggression. Age can also play a
role in ease in engaging in verbal persuasion, be-
cause we found that older children reported higher
self-efcacy in this dimension, which could be
inuenced by the increase in verbal skills that takes
place as children grow older.
The other variable which discriminates between
bullies and victims was sadness. Victims were the
only ones to report feeling sad when something
unpleasant happened. They feel unable to cope with
unpleasant situations and can count on only a few
friends, if any (Hodges, Malone, & Perry, 1997; Pel-
legrini et al., 1999), which may increase their feeling
of sadness. However, what we do not knowis whether
victims are more easily saddened than others, and
therefore more likely to be victimized, or whether the
fact that they are victims makes them more prone to
sadness. It may be that for victims sadness is a sec-
ondary emotion which follows from their vulnerabil-
ity, helplessness and incapacity to deal properly with
others aggression, which, in their turn, may even
increase their own reactive aggression.
Although we did not nd that bullies would
expect more positive outcomes from aggression, the
characteristics of the target were nevertheless
important. Children thought that both aggressive
and prosocial behavior would further their aims
more often when the target was friendly than when
he/she was aggressive. This outcome may reveal a
common tendency to consider aggressive children
as very difcult peers with whom to interact,
independent of the type of interaction (i.e., aggress-
ive or prosocial).
Defenders, outsiders and children not involved
were all low on aggression (reactive and proactive).
They did not meet aggression with aggression and
did not use it to achieve their aims. They process
social information in every step without using
aggression. They do not make hostile attributions
and, probably as a consequence, do not select anti-
social goals (e.g., retaliation), nor do they express
anger or sadness. Defenders are active children who
stand up for those who are weaker and victimized;
they are likely to be well adjusted and popular in the
peer group (Salmivalli, Lagerspetz, et al., 1996).
What is more, outsiders seem to have a capacity to
avoid harassment and to develop an adjusted cog-
nitive and emotional path. Further research is nee-
ded to uncover the characteristics which enable
outsiders and children who are not involved to
194 Marina Camodeca and Frits A. Goossens
remain detached from the bullying situation. It
would be useful for the victims to learn such skills.
To summarize, we indeed found a similarity be-
tween bullies and victims in their cognitions and
emotions (Camodeca et al., 2003; Karatzias et al.,
2002), which may be due to their common reactive
aggression and which is in line with those studies
claiming that victimization is more closely related to
externalizing problems and to aggression than to
internalizing problems (Khatri, Kupersmidt, & Pat-
terson, 2000; Kochenderfer, & Ladd, 1997; Schwartz
et al., 1998). Further studies are needed to lend
support to the nding that victims are probably not
always or not exclusively shy and withdrawn as
commonly thought, but that they may be reactively
aggressive.
Some limitations to this study can be identied. It
would be advisable to investigate whether the asso-
ciations found are also valid for other age groups
and whether such associations can be causally
interpreted (for example, whether being reactively
aggressive increases the chances of becoming a bully
or a victim, or if it is the other way round). Of course,
other causes may be found. Further research is
needed to generalize the results and to delve deeper
into SIP steps and emotions, such as the moral
emotions of shame and guilt which may inuence the
behavior of children in the group (Menesini et al.,
2003; Olthof, & Goossens, 2003). Also, the role of
follower needs further investigation, because it seems
that followers display the same types of aggression as
bullies (i.e., reactive and proactive), but their social
information processing is more similar to that of
children not involved than of bullies. In this case an
informant bias may be occurring: teachers may
overestimate the aggression of followers, or, alter-
natively, followers themselves could lack an objective
perception of themselves. Finally, since the PRS as
used by us does not assess the role of bully/victim,
including such a category could yield more interest-
ing results.
However, we think the results of the present study
contribute to the research into bullying, since new
perspectives on cognitions and emotions have been
advanced. The use of SIP, enhanced by the integra-
tion of emotions, and through its circular formula,
proved to be a reliable theoretical model to describe
the inuence of early decits on later steps. Applying
the SIP model to the topic of bullying, focusing also
on emotions, is of great importance in revealing the
way of thinking, feeling and consequently behaving
of bullies and victims and in intervening where
biases occur.
Acknowledgements
We are grateful to (in alphabetical order) Hans
M. Koot, Mark Meerum Terwogt and Carlo Schuengel
for their comments and suggestions on an earlier
draft of this paper. We are also indebted to the stu-
dents who assisted us in data collection and to the
children and teachers who took part in this study.
Correspondence to
Marina Camodeca, Faculty of Psychology and Edu-
cation, Vrije Universiteit, van der Boechorststraat 1,
1081 BT, Amsterdam, The Netherlands, Tel:
0031-20-4448979; Fax: 0031-20-4448745; Email:
M.Camodeca@psy.vu.nl.
References
Arsenio, W.F., & Lemerise, E.A. (2001). Varieties of
childhood bullying: Values, emotion processes,
and social competences. Social Development, 10,
5973.
Bjorkqvist, K., Lagerspetz, K.M.J., & Kaukiainen, A.
(1992). Do girls manipulate and boys ght? Develop-
mental trends in regard to direct and indirect aggres-
sion. Aggressive Behavior, 18, 117127.
Boulton, M.J., & Underwood, K. (1992). Bully/victim
problems among middle school children. British
Journal of Educational Psychology, 62, 7387.
Boulton, M.J., & Smith, P.K. (1994). Bully/victim
problems among middle school children: Stability,
self-perceived competence, peer perception and peer
acceptance. British Journal of Developmental Psycho-
logy, 12, 315329.
Camodeca, M., Goossens, F.A., Meerum Terwogt, M., &
Schuengel, C. (2002). Bullying and victimization
among school-age children: Stability and links to
proactive and reactive aggression. Social Develop-
ment, 11, 332345.
Camodeca, M., Goossens, F.A., Schuengel, C., &
Meerum Terwogt, M. (2003). Links between social
information processing in middle childhood and
involvement in bullying. Aggressive Behavior, 29,
116127.
Champion, K.M. (2001, April). Victims responses to
anger-provoking events: Ill get you for that! Paper
presented at the Biennial Meeting of the Society for
Research in Child Development, Minneapolis, MN.
Crick, N.R., & Dodge, K.A. (1994). A review and
reformulation of social information-processing mech-
anisms in childrens social adjustment. Psychological
Bulletin, 115, 74101.
Crick, N.R., & Dodge, K.A. (1996). Social information-
processing mechanisms in reactive and proactive
aggression. Child Development, 67, 9931002.
Crick, N.R., & Dodge, K.A. (1999). Superiority in the eye
of the beholder: A comment on Sutton, Smith and
Swettenham. Social Development, 8, 128131.
Crick, N.R., & Grotpeter, J.K. (1995). Relational aggres-
sion, gender, and social-psychological adjustment.
Child Development, 66, 710722.
Dodge, K.A. (1986). A social information processing
model of social competence in children. In M. Perl-
mutter (Ed.), The Minnesota Symposium on Child
Psychology, 18, (pp. 77125). Hillsdale, NJ: Erlbaum.
Dodge, K.A. (1991). The structure and function of
reactive and proactive aggression. In D.J. Pepler, &
Bullying, emotion and social cognition 195
K.H. Rubin (Eds.), The development and treatment of
childhood aggression (pp. 201218). Hillsdale, NJ:
Erlbaum.
Dodge, K.A., & Coie, J.D. (1987). Social-information-
processing factors in reactive and proactive aggres-
sion in childrens playgroups. Journal of Personality
and Social Psychology, 53, 11461158.
Dodge, K.A., & Crick, N.R. (1990). Social information-
processing bases of aggressive behavior in children.
Personality and Social Psychology Bulletin, 15, 822.
Dodge, K.A., Lochman, J.E., Harnish, J.D., Bates, J.E.,
& Pettit, G.S. (1997). Reactive and proactive aggres-
sion in school children and psychiatrically impaired
chronically assaultive youth. Journal of Abnormal
Psychology, 106, 3751.
Dodge, K.A., Murphy, R.R., & Buchsbaum, K. (1984).
The assessment of intention-cue detection skills in
children: Implications for developmental psycho-
pathology. Child Development, 55, 163173.
Egan, S.K., & Perry, D.G. (1998). Does low self-regard
invite victimization? Developmental Psychology, 34,
299309.
Eisenberg, N., & Fabes, R.A. (1992). Emotion regulation
and the development of social competence. M.S.
Clark (Ed.), Emotion and social behavior: Vol. 14.
Review of personality and social psychology (pp. 119
150). Newbury Park, CA: Sage.
Graham, S., & Hoehn, S. (1995). Childrens under-
standing of aggression and withdrawal as social
stigmas: An attributional analysis. Child Develop-
ment, 66, 11431161.
Graham, S., Hudley, C., & Williams, E. (1992). Attribu-
tional and emotional determinants of aggression in
African-American and Latino early adolescents. De-
velopmental Psychology, 28, 731740.
Graham, S., & Juvonen, J. (1998). Self-blame and peer
victimization in middle school: An attributional
analysis. Developmental Psychology, 34, 587599.
Hawker, D.S.J., & Boulton, M.J. (2000). Twenty years
research on peer victimization and psychosocial mal-
adjustment: A meta-analytic review of cross-sectional
studies. Journal of Child Psychology and Psychiatry,
41, 441455.
Hodges, E.V.E., Malone, M.J., & Perry, D.G. (1997).
Individual risk and social risk as interacting determin-
ants of victimization in the peer group. Developmental
Psychology, 33, 10321039.
Karatzias, A., Power, K.G., & Swanson, V. (2002).
Bullying and victimisation in Scottish secondary
schools: Same or separate entities? Aggressive Beha-
vior, 28, 4561.
Khatri, P., Kupersmidt, J.B., & Patterson, C. (2000).
Aggression and peer victimization as predictors of
self-reported behavioral and emotional adjustment.
Aggressive Behavior, 26, 345358.
Kochenderfer, B.J., & Ladd, G.W. (1997). Victimized
childrens responses to peers aggression: Behaviors
associated with reduced versus continued victimiza-
tion. Development and Psychopathology, 9, 5973.
Lemerise, E.A., & Arsenio, W.F. (2000). An integrated
model of emotion processes and cognition in social
information processing. Child Development, 71, 107
118.
Loeber, R., & Coie, J. (2001). Continuities and discon-
tinuities of development, with particular emphasis on
emotional and cognitive components of disruptive
behavior. In J. Hill, & B. Maugham (Eds.), Conduct
disorders in childhood and adolescence (pp. 379
406). Cambridge: Cambridge University Press.
Menesini, E. (1999). Bullying and emotions. In Nature
and Prevention of Bullying [On line]. Retrieved 1999
from http://www.gold.ac.uk/tmr/reports/aim2_
renze2.html.
Menesini, E., & Gini, G. (2000). Il bullismo come
processo di gruppo. Adattamento e validazione del
Questionario Ruoli dei Partecipanti alla popolazione
italiana [Bullying as a group process. Adaptation and
validation of the Participant Role Scales for Italian
population]. Eta` Evolutiva, 66, 1832.
Menesini, E., Sanchez, V., Fonzi, A., Ortega, R.,
Costabile, A., & LoFeudo, G. (2003). Moral emotions
and bullying. A cross-national comparison of differ-
ences between bullies, victims and outsiders. Aggres-
sive Behavior. 29, 515530.
Murphy, B.C., & Eisenberg, N. (1997). Young childrens
emotionality: Regulation and social functioning and
their responses when they are targets of a peers
anger. Social Development, 6, 1836.
Olthof, T., & Goossens, F.A. (2003, April). Emotional
and motivational correlates of playing a particular
role in bullying. In D.F.H. Wolke, & F.A. Goossens
(Chairs), Bullying: Data from 5 continents. Sympo-
sium conducted at the biennial meeting of the
Society for Research in Child Development, Tampa,
Florida.
Olweus, D. (1991). Bully/victim problems among
schoolchildren: Basic facts and effects of a school-
based intervention program. In D. Pepler, & K.H.
Rubin (Eds.), The development and treatment of
childhood aggression (pp. 411448). Hillsdale, NJ:
Erlbaum.
Olweus, D. (1993). Bullying at school. What we know
and what we can do. Oxford: Blackwell.
Orobio de Castro, B. (2000). Social information process-
ing and emotion in antisocial boys. Unpublished
doctoral dissertation, Free University, Amsterdam,
The Netherlands.
Oude Nijhuis, M. (2001). Pesten als groepsproces. Een
onderzoek naar validering van de Participant Role
Scales. [Bullying as a group process. A study into the
validity of the Participant Role Scales]. Unpublished
masters thesis, Free University, Amsterdam, The
Netherlands.
Pakaslahti, L. (2000). Childrens and adolescents
aggressive behavior in context: The development
and application of aggressive problem-solving strat-
egies. Aggressive and Violent Behavior, 5, 467490.
Paladino, J., & Kochenderfer-Ladd, B. (2001, April).
Coping strategies of peer-victimized children: Mod-
erators of psycho-emotional adjustment? In M. Prin-
stein, & B. Kochenderfer-Ladd (Chairs), Childrens
response to peer victimization: Moderators of psycho-
logical adjustment. Symposium conducted at the
meeting of the Society for Research in Child Develop-
ment, Minneapolis, MN.
Pellegrini, A.D., Bartini, M., & Brooks, F. (1999). School
bullies, victims, and aggressive victims: Factors
relating to group afliation and victimization in early
adolescence. Journal of Educational Psychology, 91,
216224.
196 Marina Camodeca and Frits A. Goossens
Perry, D.G., Kusel, S.J., & Perry, L.C. (1988). Victims of
peer aggression. Developmental Psychology, 24, 807
814.
Perry, D.G., Perry, L.C., & Rasmussen, P. (1986).
Cognitive social learning mediators of aggression.
Child Development, 57, 700711.
Pettit, G.S., Polaha, J.A., & Mize, J. (2001). Perceptual
and attributional processes in aggression and con-
duct problems. In J. Hill, & B. Maugham (Eds.),
Conduct disorders in childhood and adolescence (pp.
292319). Cambridge: Cambridge University Press.
Price, J.M., & Dodge, K.A. (1989). Reactive and
proactive aggression in childhood: Relations to peer
status and social context dimensions. Journal of
Abnormal Child Psychology, 17, 455471.
Pulkkinen, L. (1996). Proactive and reactive aggression
in early adolescence as precursors to anti and proso-
cial behavior in young adults. Aggressive Behavior,
22, 241257.
Quiggle, N.L., Garber, J., Panak, W.F., & Dodge, K.A.
(1992). Social information processing in aggressive
and depressed children. Child Development, 63,
13051320.
Roland, E., & Idse, T. (2001). Aggression and bullying.
Aggressive Behavior, 27, 446462.
Rubin, K.H., Bream, L.A., & Rose-Krasnor, L. (1991).
Social problem solving and aggression in childhood.
In D.J. Pepler, & K.H. Rubin (Eds.), The development
and treatment of childhood aggression (pp. 219248).
Hillsdale, NJ: Erlbaum.
Salmivalli, C. (2001). Feeling good about oneself, being
bad to others? Remarks on self-esteem, hostility, and
aggressive behavior. Aggressive and Violent Behavior,
6, 375393.
Salmivalli, C., Huttunen, A., & Lagerspetz, K.M.J.
(1997). Peer networks and bullying in schools.
Scandinavian Journal of Psychology, 38, 305312.
Salmivalli, C., Karhunen, J., & Lagerspetz, K.M.J.
(1996). How do the victims respond to bullying?
Aggressive Behavior, 22, 99109.
Salmivalli, C., Kaukiainen, A., Kaistaniemi, L., &
Lagerspetz, K.M.J. (1999). Self-evaluated self-esteem,
peer-evaluated self-esteem, and defensive egotism as
predictors of adolescents participation in bullying
situations. Personality and Social Psychology
Bulletin, 25, 12681278.
Salmivalli, C., Lagerspetz, K., Bjorkqvist, K., O

sterman,
K., & Kaukiainen, A. (1996). Bullying as a group
process: Participant roles and their relations to social
status within the group. Aggressive Behavior, 22,
115.
Salmivalli, C., Lappalainen, M., & Lagerspetz, K.M.J.
(1998). Stability and change of behavior in connection
with bullying in schools: A two-year follow-up.
Aggressive Behavior, 24, 205218.
Salmivalli, C., & Nieminen, E. (2002). Proactive and
reactive aggression among school bullies, victims and
bully-victims. Aggressive Behavior, 28, 3044.
Schippell, P., & Vasey, M.W. (2001, April). Inated
perceived self-competence and aggressive behavior in
youth. Poster presented at the meeting of the Society
for Research in Child Development, Minneapolis,
MN.
Schwartz, D., McFayden-Ketchum, S.A., Dodge, K.A.,
Pettit, G.S., & Bates, J.E. (1998). Peer group victim-
ization as a predictor of childrens behavior problems
at home and in school. Development and Psycho-
pathology, 10, 8799.
Smith, P.K. (1991). Hostile aggression as social skills
decit or evolutionary strategy? Behavioral and Brain
Sciences, 14, 315316.
Smith, P.K., & Brain, P.F. (Eds.). (2000). Bullying in
schools [Special issue]. Aggressive Behavior, 26.
Smith, P.K., Morita, Y., Junger-Tas, J., Olweus, D.,
Catalano, R.F., & Slee, P. (1999). The nature of school
bullying. A cross-national perspective. London:
Routledge.
Sutton, J. (2001). Bullies: Thugs or thinkers? The
Psychologist, 14, 530534.
Sutton, J., & Smith, P.K. (1999). Bullying as a group
process: An adaptation of the participant role
approach. Aggressive Behavior, 25, 97111.
Sutton, J., Smith, P.K., & Swettenham, J. (1999a).
Bullying and theory of mind: A critique of the social
skill decit view of anti-social behaviour. Social
Development, 8, 117127.
Sutton, J., Smith, P.K., & Swettenham, J. (1999b).
Social cognition and bullying: Social inadequacy or
skilled manipulation? British Journal of Developmen-
tal Psychology, 17, 435450.
Weiner, B. (1995). Judgments of responsibility. A foun-
dation for a theory of social conduct. New York: The
Guilford Press.
Wolke, D., Woods, S., Bloomeld, L., & Karstadt, L.
(2000). The association between direct and relational
bullying and behaviour problems among primary
school children. Journal of Child Psychology and
Psychiatry, 41, 9891002.
Zelli, A., Dodge, K.A., Lochman, J.E., Laird, R.D., &
Conduct Problems Prevention Research Group.
(1999). The distinction between beliefs legitimizing
aggression and deviant processing of social cues:
Testing measurement validity and the hypothesis that
biased processing mediates the effects of beliefs on
aggression. Journal of Personality and Social Psycho-
logy, 77, 150166.
Manuscript accepted 6 February 2004
Bullying, emotion and social cognition 197